Enzymatic Synthesis of 1-Sinapoylglucose from Free Sinapic Acid and UDP-Glucose by a Cell-Free System from Raphanus sativus Seedlings Dieter Strack Botanisches Institut der Universität zu Köln, Gyrhofstr. 15, D-5000 Köln 41

Z. Naturforsch. 35 c, 204-208 (1980); received December 28, 1979

Raphanus, Brassicaceae, Phenylpropanoid Metabolism, Glucose Ester, Esterification. Protein extracts from seedlings of Raphanus sativus catalyze the transfer of the glucosyl moiety of UDP-glucose to the carboxyl group of phenolic acids. Enzymatic activity was determined spec- trophotometrically by measuring the increase in absorbance at 360 nm and/or by the aid of high performance liquid chromatography (HPLC). From 12 phenolic acids tested as acceptors, sinapic acid was by far the best substrate. The glu- cosyltransfer to sinapic acid has a pH optimum near 7 and requires as SH group for activity, p- Chloromercuribenzoate (PCMB) inhibits activity, which can be restored by the addition of di- thiothreitol (DTT). The formation of 1-sinapoylglucose was found to be a reversible reaction, sin­ ce the addition of UDP results in a breakdown of the ester.

Introduction Materials and Methods Higher plants contain a large number of various Plant material and culture conditions are described hydroxycinnamoyl esters. Most of them might be elsewhere [10]. synthesized via hydroxycinnamoyl-CoA thiolesters as the activated reaction partners [1]. The widely Thin-layer chromatography occurring glucose esters [2], however, might be Phenolic acid derivatives were chromatographed exclusively synthesized from free hydroxycinnamic on microcrystalline cellulose (Avicel) in CAW, chlo­ acids and UDP-glucose [3, 4]. The formation of roform - acetic acid - water (3 : 2, water saturated) glucose esters of benzoic acids might follow the same and were detected under UV with and without NH3- mechanism [5], vapor. Sugar chromatography was carried out on Whereas the involvement of UDP-glucose is well the same layer with BPW, w-butanol — pyridine — established in the formation of phenylpropanoid water (3:1:1) with an overflow technique: The glycosides, containing hydroxycinnamyl alcohols [6], upper margin of the plates was connected with chro­ flavonols [7, 8], or anthocyanidins as aglyca [9], there matographic paper, which hung over the back of the is a lack of knowledge of the enzymatic reaction of plates and allowed development for 15 —20 h. Sugar glucose ester formation. detection was achieved with ammoniacal silver ni­ Seedlings of Raphanus sativus accumulate a high trate [11]. amount of 1-sinapoylglucose (80—100 nmols per cotyledon) in early stages of germination. The sinapic High performance liquid chromatography acid of this ester is derived from sinapoylcholine (sinapine), which is hydrolyzed by a specific sinapine The liquid chromatograph used was obtained esterase [10], from Spectra Physics (Santa Clara, Cal., USA) and The present report describes the enzymatic forma­ is described elsewhere [12]. tion of 1-sinapoylglucose via uridine diphosphate-D- The applied column (250 x 4 mm) was prepacked glucose (UDPG) and free acid by protein extracts with LiChrosorb RP-8 (5 |im) (Merck, Darmstadt, from young seedlings of Raphanus sativus. This reac­ Germany). Separations were accomplished by gradi­ tion is freely reversible, which might be of physio­ ent elution: in 35 min linear from 1% phosphoric logical significance for the Raphanus seedling. acid (A) to 70% methanol (B) in A + B. The flow- rate was 1 ml/min and detection at 330 or 265 nm Reprint requests to Dr. D. Strack. was achieved with a Schoeffel SF 770 UV-VIS 0341-0382/80/0300-0204 $01.00/0 detector (Kratos Inc., Trappenkamp, Germany). D. Strack • Enzymatic Synthesis of 1-Sinapoylglucose from Free Sinapic Acid 205

HPL-radiochromatograms from enzyme assays (1 n NaOH, room temperature for 30 min) and by containing UDP-[3H]glucose were obtained by frac­ analysis of the sugar [13], tionating eluants into 1 ml samples, which were col­ Tentative identification of glucose esters of other lected at the detector flow-cell outlet into 5 ml phenolic acids were done by HPLC and mild alka­ Unisolve I (Zinsser) in scintillation vials. Counting line hydrolysis. was done with a Tricarb-liquid-spectrometer (Pack­ ard). Results Substrates Identification of reaction products Phenolic acids were purchased from Fluka, Neu- In the presence of UDP-glucose protein extracts of Ulm, Merck, Darmstadt, and Roth, Karlsruhe. Raphanus sativus cotyledons are capable to form Uridine diphosphate-D-glucose (UDPG) disodium glucose esters with various phenolic acids (Table I). salt was obtained from Merck; UDP-D-[6-3H]glucose In each assay only one product was found (TLC, ammonium salt (3.1 Ci/mmol) from Amersham Buch- HPLC), exhibiting high sensitivity towards alkaline ler, Braunschweig; ADP, CDP, GDP, and UDP treatment. from Boehringer, Mannheim. When the enzyme assay contained sinapic acid, the product formed showed TLC and HPLC identity Preparation o f the cell-free extract with 1-sinapoylglucose, which was isolated from Raphanus sativus seedlings [13]. The sugar released To 50 seedlings 0.5 to 1.0 g insoluble Polyclar AT by alkali could not be distinguished from authentic (Serva, Heidelberg) was added and treated with an glucose. Incubations containing in addition UDP- Ultra Turrax homogenizer in an ice bath for 2 min [3H]glucose gave 1-sinapoylglucose with radioactivity in 10 ml potassium phosphate buffer (100 mM, found exclusively in the glucose moiety. HPLC pH 7.0), containing 10 mM dithiothreitol (DTT) analysis of this assay gave the expected simultaneous (Serva, Heidelberg). The suspension was centrifuged elution of the UV absorbing 1-sinapoylglucose and at 3,000 x g for 15 min, the filtered supernatant was its radioactivity (Fig. 1). HPLC analyses of incuba­ centrifuged at 38,000 x g for 30 min, and was finally tions, which contained other phenolic acids (Ta­ subjected to a Sephadex G-25 centrifugation. ble I), gave the same results. When these samples were treated with 1 N NaOH for 30 min at room Enzyme assay and determination of activity temperature the UV absorbing- and the 3H-product 100 |il protein extract was mixed with 100 jj.1 phe­ nolic acid solution (in buffer) and 2.7 ml potassium phosphate buffer (100 mM, pH 7.0, 10 mM DTT) and Table I. Specificity of the transfer of the glucosylmoiety of UDP-glucose to the carboxyl group of different phenolic preincubated for 30 min at 30 °C. The reaction was compounds. Activities were determined by calculating the started by the addition of 10 |il UDPG solution (in amount of label incorporation from UDP-[ 3H]glucose into water). Substrates were in a final concentration of each substrate, obtained by HPLC (see fig. 1).

1 mM each. Radioactive assays contained in addition Substrate (1 mM ) % Relative 3 p-Ci UDP-[6-3H]glucose. activity (compared to Enzyme activity was determined spectrophoto- sinapic acid= 1 0 0 ) metrically by measuring the increase in absorbance at 360 nm (PMQ II, Zeiss) and/or by high perfor­ Sinapic acid 1 0 0 D-Coumaric acid 40 mance liquid (radio)chromatography (Fig. 1 and 2). Ferulic acid 38 The absorbance of the produced 1-sinapoylglucose Caffeic acid 38 in the reaction mixture followed the Lambert-Beer 3,5-Dimethoxycinnamic acid 33 Cinnamic acid 16 law, which was determined by HPLC. Benzoic acid 39 Vanillic acid 36 Product identification Syringic acid 35 Anthranilic acid 31 Identification of 1-sinapoylglucose was achieved 4-Hydroxybenzoic acid 31 by TLC [13], HPLC [14], mild alkaline hydrolysis 3,5-Dihydroxybenzoic acid 5 206 D. Strack • Enzymatic Synthesis of 1-Sinapoylglucose from Free Sinapic Acid

0 10 20 30 0 10 20 30 Retention time (min) Retention time (min)

Fig. 1. HPLC analyses of a protein extract from Raphanus sativus cotyledons, incubated with sinapic acid and UDP- glucose and UDP-[ 3H]glucose, (a) at t 0 and (b) after 3 h incubation. 20 nl of the incubation medium was directly injected without any pretreatment.

peaks were no longer detectable. The identity of the factor of three, compared with assays without DTT. products />-coumaroyl-, feruloyl-, and caffeoylglu- 72 and 62% of this activity was obtained with 1 mM cose was further substantiated by the HPLC analysis and 100 mM DTT, respectively. PCMB at 10~3m described in ref. [14]. gave 100% inhibition, but activity could be restored up to 50% by the addition of 10 mM DTT. The Characterization of the enzymatic formation influence of 10-4 M PCMB, which inhibited to 24%, o f 1-sinapoylglucose could be neutralized by 10 mM DTT. Extraction and assay buffers with molarities be­ The amount of 1-sinapoylglucose formed from tween 50 and 100 mM potassium phosphate of pH sinapic acid and UDP-glucose was found to be values around 7.0 gave maximal activities. The addi­ proportional to added enzyme — 25 to 150^1 in tion of DTT or 2-mercaptoethanol markedly increas­ assays of 3 ml were tested - and to time up to ed the rate of sinapoylglucose synthesis. The optimal 15 min. After 3 h around 10% of the incubated free concentration of 10 mM DTT stimulated activity by a sinapic acid was converted to 1-sinapoylglucose. D. Strack ■ Enzymatic Synthesis of 1-Sinapoylglucose from Free Sinapic Acid 207

Discussion In the metabolism of plant phenolic acid esters, the central role of cinnamoyl-CoA derivatives is well established [1], However, in a few studies the in­ volvement of UDP-glucose in the formation of the most common phenolic glucose esters [3 — 5] has been demonstrated. The present report is in agree­ ment with the results of these authors and supports the assumption that this reaction might be of general occurrence for biosyntheses of phenolic glucose esters.

From the fact that a wide variety of plants is able Time (hrs) to form glucose esters from fed free phenolic acids Fig. 2. Time-course of the formation of 1-sinapoylglucose by a protein extract from Raphanus sativus cotyledons, [15], even plants which do not contain these esters incubated with sinapic acid and UDP-glucose. The forma­ naturally, one might assume, that this biochemical tion of 1 -sinapoylglucose is measured by the increase in mechanism were an unspecific detoxification reac­ absorbance at 360 nm. Arrows indicate the time where UDPG and free nucleotides in 4 independent experiments tion. were added. The present paper does not exclude this possibility for the described enzymatic formation of 1 -sinapoyl­ The initial rate of enzymatic activity was maximal glucose with Raphanus, since all applied phenolic at 40 °C; 25 °C gave only 50% of the maximal acids were found to be acceptors of the glucosyl activity and at 45 °C 80% was observed. moiety of UDP-glucose in the formation of glucose The enzymatic activity had no requirement for the esters (Table I). However, I would like to point out, metal ions Mg2+ or Ca2+, and EDTA (Titriplex II) that sinapic acid, which is the naturally occurring showed no influence on activity. main phenolic compound in Raphanus, was by far The enzymatic formation of 1-sinapoylglucose is the best acceptor. In this connection it is interesting stopped and the ester split to more than 50%, when to note, that Comer and Swain [3] found exclusively UDP (1 mM) is added to an ongoing reaction. ADP, 1 -sinapoylglucose, when incubating the common free CDP, or GDP showed no such effect (Fig. 2). hydroxycinnamic acids and UDPG with an acetone powder from Brassica oleracea, which is related to Substrate specificity Raphanus. The possible specificity of the Raphanus transferase reaction for sinapic acid is substantiated Of the phenolic acids tested for enzymatic incor­ by the fact, that highest cell-free glucosyltransfer to poration into a glucose ester, the protein preparation sinapic acid was found in those stages of germina­ exhibited high specificity towards sinapic acid (Ta­ tion, in which highest accumulation of sinapoyl­ ble I). Other naturally occurring phenolic acids were glucose can be observed [13, 16]. esterified to about 30 to 40%, compared to sinapic acid. At present we cannot decide whether contami­ Localization of the activity with sinapic acid in the nating glucosyltransferase activities with different seedling organs showed that 90% was found in substrate specificities are responsible for the observ­ cotyledons. Mixing experiments excluded possible ed capability to transfer glucose to the carboxyl inhibition factors. group of different phenolic acids or whether the Maximal glucosyltransfer activity with sinapic acid various acids esterified by the same enzyme. It was found in cotyledons of 2 day-old seedlings. At might be possible that in Raphanus cotyledons there this germination stage sinapoylglucose is accumulat­ is a base level of an unspecific transferase activity ed at its highest rate [13]. The dry seed showed 32% which is superimposed by a specific UDPG: sinapic and 4 day-old seedlings 24% activity, compared to acid glucosyltransferase, with the function to cata­ the maximally extractable one. Mixing experiments lyze the reesterification of sinapoylcholine to sinapoyl­ excluded possible inhibitory effects. glucose. The possible involvement of a second major 208 D. Strack • Enzymatic Synthesis of 1-Sinapoylglucose from Free Sinapic Acid sinapoyl derivative [16] in this reesterification reac­ in a following reesterification reaction resulting in tion is proven to be unlikely [17]. the accumulation of sinapoylmalate [13, 17]. Since all The results presented here suggest that the glu- attempts failed to demonstrate at an appropriate cosyltransferase reaction with sinapic acid is freely germination stage an esterase activity acting on sina- reversible in vitro and it might be that UDPG is polyglucose, the release of free sinapic acid from this formed from 1-sinapoylglucose and UDP. Surprisingly compound could be entirely dependent on th re­ also a free reversibility of a UDPG: flavonol 3-0- versibility of the reaction described in this paper. glucosyltransferase reaction is described by Sutter However, it cannot be excluded that the free form of and Grisebach [18]. A free reversibility of the UDPG: sinapic acid is bypassed by a direct transacylation sinapic acid glucosyltransferase activity — if oc­ to form sinapoylmalate [19]. curring in situ — might be of physiological signifi­ Experiments are underway to clarify the points cance for Raphanus in two ways. Firstly it could discussed above. provide the cotyledon with large amounts of UDPG. Secondly this reversibility of the sinapoylglucose Acknowledgement formation may be of importance regarding the The support by the Deutsche Forschungsgemein­ availability of free sinapic acid. The acid is needed schaft is gratefully acknowledged.

[1] M. H. Zenk, Biochemistry of plant phenolics (edit, [10] G. Nurmann and D. Strack, Z. Naturforsch. 34 c, 715 by T. Swain, J. B. Harbome, and C. F. Sumere), Vol. (1979). 12,139, Plenum Press, New York 1979. [11] E. C. Bate-Smith and R. G. Westall, Biochim. Bio­ [2] J. B. Harbome, Biochemistry of phenolic compounds phys. Acta 4,422 (1950). (edit, by J. B. Harbome), p. 129, Academic Press, [12] D. Strack, G. B. Feige, and R. Kroll, Z. Naturforsch. London 1964. 34 c, 6 7 9 (1 9 7 9 ). [3] J. J. Comer and T. Swain, Nature 207,634 (1965). [13] D. Strack, Z. Pflanzenphysiol. 8 4 , 139 (1977). [4] J.-J. Macheix, C. R. Acad. Sc.284,D-33 (1977). [14] J. Krause and D. Strack, J. Chromatogr. 176, 4 65 [5] G. Jacobelli, M. J. Tabone, and D. Tabone, Bull. Soc. (1 9 7 9 ). Chim. Biol. 40,955 (1958). [15] J. B. Harbome and J. J. Comer, Biochem. J. 81, 2 42 [6 ] R. K. Ibrahim and H. Grisebach, Arch. Biochem. Bio­ (1 961). phys. 176,700 (1976). [16] D. Strack, N. Tkotz, and M. Klug, Z. Pflanzenphysiol. [7] A. Sutter, R. Ortmann, and H. Grisebach, Biochim. 89,343(1978). Biophys. Acta 258,71 (1972). [17] D. Strack and M. Linscheid, in preparation. [8 ] J. L. Shute, P. S. Jourdan, and R. L. Mansell, Z. Na- [18] A. Sutter and H. Grisebach, Arch. Biochem. Biophys. turforsch. 34 c, 738 (1979). 167,444(1975). [9] N. A. M. Saleh, H. Fritsch, P. Witkop, and H. Grise­ [19] R. Schlepphorst and W. Barz, Planta Med. 36, 333 bach, Planta 133,41 (1976). (1 9 7 9 ).